Use “Mark as Suspicious” before enabling “Reject”
Singular’s Fraud Prevention Suite provides powerful tools that affect attribution decisions in real time. Both Singular’s Prevention Methods and the user’s custom rules might run into edge-cases which produce unwanted behaviour in the form of false-positives or false-negatives.
Therefore the best practice for users leveraging Singular’s Advanced Fraud Prevention tools is to start gradually:
- Use “Mark as Suspicious” to test the rules on your data before switching to “Reject”. This will allow you to monitor the performance of these rules in the Suspicious Report without risking changes to your ongoing marketing campaigns or negatively impact your volume unexpectedly.
- Once the rule has been tested, you should promote it to actively “Reject” to prevent CPI/CPA billing postbacks from being sent.
For example, you might find that most of your end-users come from blacklisted IPs because you target end users that use VPN services.
Leverage Custom Rules
Each application has its own fraud consideration therefore we recommend using the custom rules to fit Singular Fraud Prevention to your specific use-cases for example:
- Whitelist partners that are guaranteed to not fraud you. For example internal crosspromo.
- Use the rules to reject installs that don’t match a campaign’s targeting options.
Leverage Singular’s Alerts
Sometimes the best way to catch fraud is to look at KPIs.
Users can set alerts on specific KPIs to know when a source or a publisher performs extremely good or extremely poorly. These abnormalities hint at possible fraud.
After being alerted, the user may choose to set a rejection for the specific publisher, which can be done using the Fraud Prevention’s Custom User Rules.